Credal Graph Neural Networks for Robust Uncertainty Quantification

ICLR 2026 Conference Submission19801 Authors

Keywords: graph neural networks, credal learning, uncertainty quantification

TL;DR: We introduce credal graph neural networks (CGNNs), the first GNNs that output set-valued predictions to better capture epistemic uncertainty, and show they achieve state-of-the-art robustness to distributional shift on heterophily graphs.

Abstract: Uncertainty quantification is essential for deploying reliable Graph Neural Networks (GNNs), where existing approaches primarily rely on Bayesian inference or ensembles. In this paper, we introduce the first credal graph neural networks (CGNNs), which extend credal learning to the graph domain by training GNNs to output set-valued predictions in the form of credal sets. To account for the distinctive nature of message passing in GNNs, we develop a complementary approach to credal learning that leverages different aspects of layer-wise information propagation. We assess our approach on uncertainty quantification in node classification under out-of-distribution conditions. Our analysis highlights the critical role of the graph homophily assumption in shaping the effectiveness of uncertainty estimates. Extensive experiments demonstrate that CGNNs deliver more reliable representations of epistemic uncertainty and achieve state-of-the-art performance under distributional shift on heterophilic graphs.

Primary Area: learning on graphs and other geometries & topologies

Submission Number: 19801